arXiv subjects

Wei Wu

Publications and source records attributed to Wei Wu.

At least 415 records · Page 23Linked to original sources

arXiv2021

Context-Aware Graph Convolution Network for Target Re-identification

Most existing re-identification methods focus on learning robust and discriminative features with deep convolution networks. However, many of them consider content similarity separately and fail to utilize the context information of the query and gallery sets, e.g. probe-gallery and gallery-gallery relations, thus hard samples may not be well solved due to the limited or even misleading information. In this paper, we present a novel Context-Aware Graph Convolution Network (CAGCN), where the probe-gallery relations are encoded into the graph nodes and the graph edge connections are well controlled by the gallery-gallery relations. In this way, hard samples can be addressed with the context information flows among other easy samples during the graph reasoning. Specifically, we adopt an effective hard gallery sampler to obtain high recall for positive samples while keeping a reasonable graph size, which can also weaken the imbalanced problem in training process with low computation complexity.Experiments show that the proposed method achieves state-of-the-art performance on both person and vehicle re-identification datasets in a plug and play fashion with limited overhead.

cs.CV↗

arXiv2021

Learning Statistical Texture for Semantic Segmentation

Existing semantic segmentation works mainly focus on learning the contextual information in high-level semantic features with CNNs. In order to maintain a precise boundary, low-level texture features are directly skip-connected into the deeper layers. Nevertheless, texture features are not only about local structure, but also include global statistical knowledge of the input image. In this paper, we fully take advantages of the low-level texture features and propose a novel Statistical Texture Learning Network (STLNet) for semantic segmentation. For the first time, STLNet analyzes the distribution of low level information and efficiently utilizes them for the task. Specifically, a novel Quantization and Counting Operator (QCO) is designed to describe the texture information in a statistical manner. Based on QCO, two modules are introduced: (1) Texture Enhance Module (TEM), to capture texture-related information and enhance the texture details; (2) Pyramid Texture Feature Extraction Module (PTFEM), to effectively extract the statistical texture features from multiple scales. Through extensive experiments, we show that the proposed STLNet achieves state-of-the-art performance on three semantic segmentation benchmarks: Cityscapes, PASCAL Context and ADE20K.

cs.CV↗

arXiv2021

BSN++: Complementary Boundary Regressor with Scale-Balanced Relation Modeling for Temporal Action Proposal Generation

Generating human action proposals in untrimmed videos is an important yet challenging task with wide applications. Current methods often suffer from the noisy boundary locations and the inferior quality of confidence scores used for proposal retrieving. In this paper, we present BSN++, a new framework which exploits complementary boundary regressor and relation modeling for temporal proposal generation. First, we propose a novel boundary regressor based on the complementary characteristics of both starting and ending boundary classifiers. Specifically, we utilize the U-shaped architecture with nested skip connections to capture rich contexts and introduce bi-directional boundary matching mechanism to improve boundary precision. Second, to account for the proposal-proposal relations ignored in previous methods, we devise a proposal relation block to which includes two self-attention modules from the aspects of position and channel. Furthermore, we find that there inevitably exists data imbalanced problems in the positive/negative proposals and temporal durations, which harm the model performance on tail distributions. To relieve this issue, we introduce the scale-balanced re-sampling strategy. Extensive experiments are conducted on two popular benchmarks: ActivityNet-1.3 and THUMOS14, which demonstrate that BSN++ achieves the state-of-the-art performance. Not surprisingly, the proposed BSN++ ranked 1st place in the CVPR19 - ActivityNet challenge leaderboard on temporal action localization task.

cs.CV↗

arXiv2021

SuperNeurons: FFT-based Gradient Sparsification in the Distributed Training of Deep Neural Networks

The performance and efficiency of distributed training of Deep Neural Networks highly depend on the performance of gradient averaging among all participating nodes, which is bounded by the communication between nodes. There are two major strategies to reduce communication overhead: one is to hide communication by overlapping it with computation, and the other is to reduce message sizes. The first solution works well for linear neural architectures, but latest networks such as ResNet and Inception offer limited opportunity for this overlapping. Therefore, researchers have paid more attention to minimizing communication. In this paper, we present a novel gradient compression framework derived from insights of real gradient distributions, and which strikes a balance between compression ratio, accuracy, and computational overhead. Our framework has two major novel components: sparsification of gradients in the frequency domain, and a range-based floating point representation to quantize and further compress gradients frequencies. Both components are dynamic, with tunable parameters that achieve different compression ratio based on the accuracy requirement and systems' platforms, and achieve very high throughput on GPUs. We prove that our techniques guarantee the convergence with a diminishing compression ratio. Our experiments show that the proposed compression framework effectively improves the scalability of most popular neural networks on a 32 GPU cluster to the baseline of no compression, without compromising the accuracy and convergence speed.

cs.DC↗

arXiv2021

Non-Markovian Sensing of a Quantum Reservoir

Quantum sensing explores protocols using the quantum resource of sensors to achieve highly sensitive measurement of physical quantities. The conventional schemes generally use unitary dynamics to encode quantities into sensor states. In order to measure the spectral density of a quantum reservoir, which plays a vital role in controlling the reservoir-caused decoherence to microscopic systems, we propose a nonunitary-encoding optical sensing scheme. Although the nonunitary dynamics for encoding in turn degrades the quantum resource, we surprisingly find a mechanism to make the encoding time a resource to improve the precision and to make the squeezing of the sensor a resource to surpass the shot-noise limit. Our result shows that it is due to the formation of a sensor-reservoir bound state. Enriching the family of quantum sensing, our scheme gives an efficient way to measure the quantum reservoir and might supply an insightful support to decoherence control.

quant-ph↗

arXiv2021

Electro-Optic Lithium Niobate Metasurfaces

Many applications of metasurfaces require an ability to dynamically change their properties in time domain. Electrical tuning techniques are of particular interest, since they pave a way to on-chip integration of metasurfaces with optoelectronic devices. In this work, we propose and experimentally demonstrate an electro-optic lithium niobate (EO-LN) metasurface that shows dynamic modulations to phase retardation of transmitted light. Quasi-bound states in the continuum (QBIC) are observed from our metasurface. And by applying external electric voltages, the refractive index of the LN is changed by Pockels EO nonlinearity, leading to efficient phase modulations to the transmitted light around the QBIC wavelength. Our EO-LN metasurface opens up new routes for potential applications in the field of displaying, pulse shaping, and spatial light modulating.

physics.optics↗

arXiv2021

BaPipe: Exploration of Balanced Pipeline Parallelism for DNN Training

The size of deep neural networks (DNNs) grows rapidly as the complexity of the machine learning algorithm increases. To satisfy the requirement of computation and memory of DNN training, distributed deep learning based on model parallelism has been widely recognized. We propose a new pipeline parallelism training framework, BaPipe, which can automatically explore pipeline parallelism training methods and balanced partition strategies for DNN distributed training. In BaPipe, each accelerator calculates the forward propagation and backward propagation of different parts of networks to implement the intra-batch pipeline parallelism strategy. BaPipe uses a new load balancing automatic exploration strategy that considers the parameters of DNN models and the computation, memory, and communication resources of accelerator clusters. We have trained different DNNs such as VGG-16, ResNet-50, and GNMT on GPU clusters and simulated the performance of different FPGA clusters. Compared with state-of-the-art data parallelism and pipeline parallelism frameworks, BaPipe provides up to 3.2x speedup and 4x memory reduction in various platforms.

cs.DC↗

arXiv2021

Magneto-Acoustic Waves in Magnetic Twisted Flux Tube

At present, many works about magneto-hydrodynamics wave diagnostics in magnetic flux tubes are based on some pioneer works not considering the contributions of magnetic twist. Other works considered the effect on magneto-hydrodynamics waves, but the dispersion relationship they presented only gave the wave modes of m=0,1,2. The kink mode of m=-1 was absent. Therefore, in this work we would like to present a complete dispersion relationship that includes both magnetic twist and the wave mode of m=-1. Analogous to the m=+1 wave mode, the mode of m=-1 also exhibits the mode change at finite kr0, from body to surface mode. The phase speeds of this mode are usually less than those of m=+1 mode. The harmonic curves of m=+1/-1 modes in dispersion relationship diagrams are approximately symmetric in respect to a characteristic velocity, e.g. the tube velocity in flux tubes. Based on the present dispersion relationship, we revisit the issue of spiral wave patterns in sunspot and find that the magnetic twist has no great influence on their morphology in the frame of linear perturbation analysis.

astro-ph.SR↗

arXiv2020

Conditional Text Generation for Harmonious Human-Machine Interaction

In recent years, with the development of deep learning, text generation technology has undergone great changes and provided many kinds of services for human beings, such as restaurant reservation and daily communication. The automatically generated text is becoming more and more fluent so researchers begin to consider more anthropomorphic text generation technology, that is the conditional text generation, including emotional text generation, personalized text generation, and so on. Conditional Text Generation (CTG) has thus become a research hotspot. As a promising research field, we find that many efforts have been paid to exploring it. Therefore, we aim to give a comprehensive review of the new research trends of CTG. We first summary several key techniques and illustrate the technical evolution route in the field of neural text generation, based on the concept model of CTG. We further make an investigation of existing CTG fields and propose several general learning models for CTG. Finally, we discuss the open issues and promising research directions of CTG.

cs.CL↗

arXiv2020

Estimation of the Laser Frequency Nosie Spectrum by Continuous Dynamical Decoupling

Decoherence induced by the laser frequency noise is one of the most important obstacles in the quantum information processing. In order to suppress this decoherence, the noise power spectral density needs to be accurately characterized. In particular, the noise spectrum measurement based on the coherence characteristics of qubits would be a meaningful and still challenging method. Here, we theoretically analyze and experimentally obtain the spectrum of laser frequency noise based on the continuous dynamical decoupling technique. We first estimate the mixture-noise (including laser and magnetic noises) spectrum up to $(2π)$530 kHz by monitoring the transverse relaxation from an initial state $+X$, followed by a gradient descent data process protocol. Then the contribution from the laser noise is extracted by enconding the qubits on different Zeeman sublevels. We also investigate two sufficiently strong noise components by making an analogy between these noises and driving lasers whose linewidth assumed to be negligible. This method is verified experimentally and finally helps to characterize the noise.

quant-ph↗

arXiv2020

Observation of $\mathcal{PT}$-symmetric quantum coherence in a single ion system

Parity-time($\mathcal{PT}$)-symmetric systems, featuring real eigenvalues despite its non-Hermitian nature, have been widely utilized to achieve exotic functionalities in the classical realm, such as loss-induced transparency or lasing revival. By approaching the exceptional point (EP) or the coalescences of both eigenvalues and eigenstates, unconventional effects are also expected to emerge in pure quantum $\mathcal{PT}$ devices. Here, we report experimental evidences of spontaneous $\mathcal{PT}$ symmetry breaking in a single cold $^{40}\mathrm{Ca}^{+}$ ion, and more importantly, a counterintuitive effect of perfect quantum coherence occurring at the EP. Excellent agreement between experimental results and theoretical predictions is identified. In view of the versatile role of cold ions in building quantum memory or processor, our experiment provides a new platform to explore and utilize pure quantum EP effects, with diverse applications in quantum engineering of trapped ions.

quant-ph↗

arXiv2020

Improving EEG Decoding via Clustering-based Multi-task Feature Learning

Accurate electroencephalogram (EEG) pattern decoding for specific mental tasks is one of the key steps for the development of brain-computer interface (BCI), which is quite challenging due to the considerably low signal-to-noise ratio of EEG collected at the brain scalp. Machine learning provides a promising technique to optimize EEG patterns toward better decoding accuracy. However, existing algorithms do not effectively explore the underlying data structure capturing the true EEG sample distribution, and hence can only yield a suboptimal decoding accuracy. To uncover the intrinsic distribution structure of EEG data, we propose a clustering-based multi-task feature learning algorithm for improved EEG pattern decoding. Specifically, we perform affinity propagation-based clustering to explore the subclasses (i.e., clusters) in each of the original classes, and then assign each subclass a unique label based on a one-versus-all encoding strategy. With the encoded label matrix, we devise a novel multi-task learning algorithm by exploiting the subclass relationship to jointly optimize the EEG pattern features from the uncovered subclasses. We then train a linear support vector machine with the optimized features for EEG pattern decoding. Extensive experimental studies are conducted on three EEG datasets to validate the effectiveness of our algorithm in comparison with other state-of-the-art approaches. The improved experimental results demonstrate the outstanding superiority of our algorithm, suggesting its prominent performance for EEG pattern decoding in BCI applications.

eess.SP↗

arXiv2020

Rainbow Perfect and Near-Perfect Matchings in Complete Graphs with Edges Colored by Circular Distance

Given an edge-colored complete graph $K_n$ on $n$ vertices, a perfect (respectively, near-perfect) matching $M$ in $K_n$ with an even (respectively, odd) number of vertices is rainbow if all edges have distinct colors. In this paper, we consider an edge coloring of $K_n$ by circular distance, and we denote the resulting complete graph by $K^{\bullet}_n$. We show that when $K^{\bullet}_n$ has an even number of vertices, it contains a rainbow perfect matching if and only if $n=8k$ or $n=8k+2$, where $k$ is a nonnegative integer. In the case of an odd number of vertices, Kirkman matching is known to be a rainbow near-perfect matching in $K^{\bullet}_n$. However, real-world applications sometimes require multiple rainbow near-perfect matchings. We propose a method for using a recursive algorithm to generate multiple rainbow near-perfect matchings in $K^{\bullet}_n$.

math.CO↗

arXiv2020

An Iterated Dual Substitution Approach for Binary Integer Programming Problems under the Min-Max Regret Criterion

We consider binary integer programming problems with the min-max regret objective function under interval objective coefficients. We propose a new heuristic framework, which we call the iterated dual substitution (iDS) algorithm. The iDS algorithm iteratively invokes a dual substitution heuristic and excludes from the search space any solution already checked in previous iterations. In iDS, we use a best-scenario-based lemma to improve performance. We apply iDS to four typical combinatorial optimization problems: the knapsack problem, the multidimensional knapsack problem, the generalized assignment problem, and the set covering problem. For the multidimensional knapsack problem, we compare the iDS approach with two algorithms widely used for problems with the min-max regret criterion: a fixed-scenario approach, and a branch-and-cut approach. The results of computational experiments on a broad set of benchmark instances show that the proposed iDS approach performs best on most tested instances. For the knapsack problem, the generalized assignment problem, and the set covering problem, we compare iDS with state-of-the-art results. The iDS algorithm successfully updates best known records for a number of benchmark instances.

math.OC↗

arXiv2020

Are Pre-trained Language Models Knowledgeable to Ground Open Domain Dialogues?

We study knowledge-grounded dialogue generation with pre-trained language models. Instead of pursuing new state-of-the-art on benchmarks, we try to understand if the knowledge stored in parameters of the pre-trained models is already enough to ground open domain dialogues, and thus allows us to get rid of the dependency on external knowledge sources in generation. Through extensive experiments on benchmarks, we find that by fine-tuning with a few dialogues containing knowledge, the pre-trained language models can outperform the state-of-the-art model that requires external knowledge in automatic evaluation and human judgment, suggesting a positive answer to the question we raised.

cs.CL↗

arXiv2020

Fast Cooling of Trapped Ion in Strong Sideband Coupling Regime

Trapped ion in the Lamb-Dicke regime with the Lamb-Dicke parameter $η\ll1$ can be cooled down to its motional ground state using sideband cooling. Standard sideband cooling works in the weak sideband coupling limit, where the sideband coupling strength is small compared to the natural linewidth $γ$ of the internal excited state, with a cooling rate much less than $γ$. Here we consider cooling schemes in the strong sideband coupling regime, where the sideband coupling strength is comparable or even greater than $γ$. We derive analytic expressions for the cooling rate and the average occupation of the motional steady state in this regime, based on which we show that one can reach a cooling rate which is proportional to $γ$, while at the same time the steady state occupation increases by a correction term proportional to $η^{2}$ compared to the weak sideband coupling limit. We demonstrate with numerical simulations that our analytic expressions faithfully recover the exact dynamics in the strong sideband coupling regime.

quant-ph↗

arXiv2020

Task Bench: A Parameterized Benchmark for Evaluating Parallel Runtime Performance

We present Task Bench, a parameterized benchmark designed to explore the performance of parallel and distributed programming systems under a variety of application scenarios. Task Bench lowers the barrier to benchmarking multiple programming systems by making the implementation for a given system orthogonal to the benchmarks themselves: every benchmark constructed with Task Bench runs on every Task Bench implementation. Furthermore, Task Bench's parameterization enables a wide variety of benchmark scenarios that distill the key characteristics of larger applications. We conduct a comprehensive study with implementations of Task Bench in 15 programming systems on up to 256 Haswell nodes of the Cori supercomputer. We introduce a novel metric, minimum effective task granularity to study the baseline runtime overhead of each system. We show that when running at scale, 100 μs is the smallest granularity that even the most efficient systems can reliably support with current technologies. We also study each system's scalability, ability to hide communication and mitigate load imbalance.

cs.DC↗

arXiv2020

Optical and magnetic signatures of the topological edge states in a dimerised donor chain

We have studied the excited states of a one-dimensional donor dimer array by using time-dependent Hartree-Fock and density-functional theories. We find that tuning the inter-donor distances can induce a topological phase transition from a topologically trivial anti-ferromagnetic ground state to a topological phase supporting spin-polarised edge states. Significant changes in the optical spectra accompany this transition, providing signatures for the phase transition and the existence of the edge state; the edge states lead to a robust $1s\rightarrow 2p$ atomic transition. By contrast, the bulk of the chain becomes spinless, inducing a reduction in the HOMO-LUMO gap and a decrease of the excitation energies in the optical spectra. The inclusion of physics beyond the Su-Schrieffer-Heeger-Hubbard model (long-range electron correlations and basis states beyond the 1s manifold) has a critical effect on the results and leads to the novel prediction of a change in optical response related to a topological phase transition. Our results provide a theoretical foundation for using artificial donor linear arrays in semiconductors to form topological edge states by design.

cond-mat.mes-hall↗